Haroon MukhtarSenior Full Stack Engineer · AI Engineer

Challenge

Equity research firms face information overload during earnings seasons, requiring analysts to manually monitor live events, cross-reference internal reports, draft daily digests, and track complex KPI anecdotes across fragmented tools and data sources.

Solution

Designed and shipped an agentic earnings-intelligence system for a financial research firm: a self-healing daily monitor that ingests earnings/event data, drafts digest emails, and updates calendars unattended, plus a 5-pass anecdote-generation pipeline and a semantic-recall research chat agent — all backed by a code-owned dedup/state layer with LLM calls scoped to pure judgment tasks.

Impact

Built a multi-agent financial research platform (TypeScript/Mastra) that automates earnings monitoring, analyst report drafting, and research chat — orchestrating LLM reasoning with deterministic code, live MCP integrations (Aiera, Gmail, Drive, Calendar), and a custom React front end.

Core features

• Self-healing daily Earnings & Events Monitor pulling live data via MCP
• Live MCP integrations with Aiera, Gmail, Google Drive, and Google Calendar
• Unattended automated synthesis of analyst-grade earnings digests and email drafts
• 5-pass LLM pipeline generating polished analyst anecdotes with verbatim KPI grounding
• Conversational research chat agent with semantic memory recall and per-request tool routing
• Custom React front end for research workflows and real-time interaction

Engineering highlights

• Mastra agent framework orchestrating LLM reasoning with deterministic TypeScript code
• Architected for reliability: LLMs scoped strictly to bounded judgment calls (structured in/out)
• Deterministic code-owned state, deduplication, and templating layer
• Zero human-in-the-loop automated daily calendar updates and email drafting pipeline

Challenge

Restaurant businesses needed professional, high-performance websites without engineering teams — and the platform behind them had to handle publishing, collaboration, SEO, accessibility, and hosting at scale.

Solution

A custom no-code builder with a drafting and templating engine, dual Live/Staging publishing workflows, real-time collaborative editing, and a scalable publishing pipeline that provisions domains and SSL automatically on AWS. Led the migration from Gatsby to Next.js to unlock a more scalable rendering architecture.

Impact

Turned website delivery into a self-serve product: restaurant sites go from draft to a live, SEO-ready, accessible production deployment without engineering involvement.

Core features

• Live / staging publishing workflows with safe promotion
• Collaborative multi-user editing environment
• Drafting & templating engine for rapid site creation
• Analytics dashboard on GA4 + Google Search Console APIs
• Media library with editing and an optimization pipeline
• Independent mobile-specific CSS override system
• Domain & SSL configuration automation on AWS

Engineering highlights

• Gatsby → Next.js migration for scalable architecture
• ADA-compliant rendering across generated sites
• Strong Core Web Vitals / Lighthouse focus on published output
• Publishing workflow designed to scale across many client sites

Challenge

Generic chatbots can talk about markets but can't act on a user's actual portfolio. The goal: an assistant that reasons over live holdings, retrieves relevant context, and executes multi-step workflows with real tools.

Solution

A LangChain-based agent with ReAct-style tool orchestration behind a FastAPI backend and a Next.js streaming chat frontend. ChromaDB powers RAG retrieval, persistent memory keeps conversations context-aware, and the agent plans multi-step reasoning chains across portfolio, pricing, and news tools.

Impact

Demonstrates end-to-end applied AI: an agent that observes, reasons, retrieves, and acts on real user data through a production-shaped architecture.

Core features

• Streaming chat UI built with Next.js
• Ollama LLM integration behind FastAPI
• Portfolio analysis & market insights
• Portfolio CRUD tools the agent invokes directly
• Coin price lookup & financial news retrieval tools
• Persistent memory with context-aware multi-step reasoning

Engineering highlights

• LangChain agent framework with ReAct-style tool orchestration
• ChromaDB vector store powering the RAG pipeline
• Deliberate separation of reasoning, retrieval, and tool execution
• Agent engineering, not a chatbot wrapper

Challenge

Growing businesses juggle purchase orders, shipments, invoices, and financial records across disconnected tools, losing visibility and duplicating work.

Solution

A modular business platform that unifies operational workflows. Domains like Inventory, Finance, and CRM live as separated modules inside a monorepo, keeping the system maintainable while workflows stay integrated end to end.

Impact

One system of record for operations — from purchase to payment — replacing fragmented spreadsheets and manual handoffs.

Core features

• Purchase orders, shipping, and invoicing workflows
• Financial workflows & reporting
• Vendor & customer management
• Activity tracking & operational automation

Engineering highlights

• Modular monorepo architecture
• Separated domains: Inventory, Finance, CRM
• Designed for maintainability as operations scale

Challenge

Remote education needed more than video calls — it needed structured scheduling, interactive teaching tools, and role-specific workflows in one platform.

Solution

A full LMS with real-time classroom infrastructure: class scheduling, video and screen sharing, collaborative whiteboarding with annotations, and live chat — plus dedicated portals for each role in the organization.

Impact

Enabled fully remote structured teaching with interactive tools that kept lessons engaging beyond a plain video call.

Core features

• Class scheduling & management
• Video conferencing & screen sharing
• Whiteboarding with live annotations
• Real-time chat
• Teacher, sales, and admin portals

Engineering highlights

• Real-time multi-user infrastructure
• Role-based product surfaces on a shared platform